Manual Background

lrichey

Hello,

I have been using a manual background for some of the trouble areas in my TOPAS fit. The function takes on this form:

Bkg_Straight_Line_First (0, X1, 5.678)
Bkg_Straight_Line (0, 1, 13.456, 7.688)
Bkg_Straight_Line (1, 2, 13.556, 12.678)
Bkg_Straight_Line (2, 3, 13.656, 16.678)
Bkg_Straight_Line (3, 4, X2, 5.678)

etc, where the first two columns are line segments relating each line to the next, the third column is the 2theta value, and the 4th column is the intensity.
This is a macro as described in http://topas.dur.ac.uk/topaswiki/doku.php?id=background_straight_line_segments

What I would like to do is allow the intensity parameter to refine so I don't have to change each and every one to make the fit work. I was thinking something like this so only the intensity refines, but I don't think this function is meant to work like this:

Bkg_Straight_Line (!nx1, !ny1, !the1, in1, 2, 3, 13.656, 16.678)

This doesn't work. Any ideas?

Thanks!

alancoelho

The y values are arguments to the macro and are used within the fit_obj equation of the macro. They can therefore be a prm; ie.

prm scale_bkg 1
Bkg_Straight_Line_First (0, X1, scale_bkg 5.678)
Bkg_Straight_Line (0, 1, 13.456, scale_bkg 7.688)
Bkg_Straight_Line (1, 2, 13.556, scale_bkg 12.678)
etc...

lrichey

Thanks! That is a lot easier.

For future users, I do find it helpful to link some of the scaling factors together and some not, just depending upon the area I am fitting

prm scale_bkg1 1.1259
prm scale_bkg2 2.1259

Bkg_Straight_Line_First (0, X1, scale_bkg1 5.678)
Bkg_Straight_Line (0, 1, 13.456, scale_bkg2 7.688)
Bkg_Straight_Line (1, 2, 13.556, scale_bkg1 12.678)

haoliu

Hi,

I have used the above method to generate a manual background. The background in my case is very complex and I have to use more than 100 points, and Topas complained "X limits invalid for fit_obj" when I run the refinement; however, after taking off a few tens of background points, the refinement runs fine. It seems there is a maximum number of fit_obj Topas can do, and I tried to work it out myself only to find this number changes from one background to another.

Does anyone know what the problem is and how to solve it?

Thanks,
Hao

alancoelho

Hao

There should be no limits on things like number of fit_obj's except for memory considerations. The "X limits invalid for fit_obj" error means that the fit_obj in question is outside of the refined data range. If you get stuck then e-mail me your INP/.data file and I will have a look

alan

silveira

Greetings all,

I am trying to do something similar but in a sequential refinement.

Every file has a different background due to phase transformation, temperature, etc.

So I would like to update the Yobs based on the actual X value of every diffraction profile like the following:

Bkg_Straight_Line_First(0 , X1, Yobs(X1) )
Bkg_Straight_Line(0 ,1 ,1.36499 , Yobs(1.36499) )
Bkg_Straight_Line(1 ,2 ,1.71453 , Yobs(1.71453) )
Bkg_Straight_Line(2 ,3 ,2.20625 , Yobs(2.20625) )
Bkg_Straight_Line(3 ,4 ,X2 , Yobs(X2) )

I was not able to understand how I could do this based on the Technical reference. I tried to use Yobs_dx_at, but the whole bkg goes to 0. Probably using this does not make any sense.

My Topas is V6, hence I don't have the Yobs_at command.

Does anyone have a tip on that?

Best regards!
Antonio

rowlesmr

One idea would be to do some preprocessing.

Look through all of your data files and extract the Yobs values you want to use as bkg points

Then do something like this in your input file:

#list filename bkg00 bkg01 bkg02 bkg03 bkg04 {
file1.xy 241 325 512 471 536
file2.xy 256 378 575 564 646
...
}

...

Bkg_Straight_Line_First(0      ,    X1,            bkg00(Run_Number) )         
Bkg_Straight_Line(0    ,1       ,1.36499    ,    bkg01(Run_Number) )
Bkg_Straight_Line(1    ,2       ,1.71453    ,    bkg02(Run_Number) )
Bkg_Straight_Line(2    ,3       ,2.20625    ,    bkg03(Run_Number) )
Bkg_Straight_Line(3    ,4       ,X2            ,    bkg04(Run_Number) )

Matthew

silveira

Thanks for your idea Matthew.

I created a Matlab script to extract the intensity of many arbitrary points in the diffractogram.

At the end of the day, I got worse results than those obtained by the Chebyshev polynomial background. <_<

It's a very basic code since I am very bad at coding, but I am posting it here in case it might help someone in the future.

Antonio